Using Multivariate Statistics to Explain Nutrient Reduction in Soil Treatment Units

Nitrogen (N) and phosphorus (P) are two nutrients that pose risk of contamination of groundwater from soil treatment units (STUs). Soil and chemical data relevant to N and P treatment were collected from 84 experiments, to determine the significance of each parameter in N and P removal. Parameter included N and P concentrations with depth in various soils, along with parameters such as soil type, hydraulic loading rate (HLR) and frequency of wastewater discharge (continuous or time-dosing), and were analyzed for variance using univariate and multivariate statistics. A reliable bivariate relationship (r2=0.891) was developed for reduction of N in clay loam soils as function of both soil depth and HLR. Univariate statistical relationships were not generally successful at yielding reliable empirical relationships to predict N and P removal for different soil types. Results from multi-factorial ANOVA test (General Linear Model) have shown that for field and laboratory experiments combined, HLR is the most important factor in determining N removal, more than soil type, depth and frequency of discharge. When field and laboratory experiments are analyzed separately, soil type becomes the greatest contributor to variability of N treatment in laboratory STU experiments, While HLR is still the most important factor in field experiments.